Twelve Labs launches Jockey, a video-editing AI agent
Jockey finds and edits scenes on spoken request, and works inside ChatGPT and Claude as well as via API.
Twelve Labs announced on October 16 the launch of Jockey, a video-understanding AI agent that lets users find and edit scenes in videos by making spoken requests. The company detailed the product in its announcement, according to aitimes.com.
Jockey applies embedding and metadata generation technology to Twelve Labs' own foundation models to locate desired scenes and carry out tasks within a video. Those tasks include creating rough cuts of highlight scenes from long source videos and pulling out only the clips that drew strong reactions.
The agent also handles social video: when a link from YouTube or TikTok is entered, Jockey analyzes trend and engagement metrics and recommends reference videos for promotion. Results can be shared with team members.
Jockey is designed to run in AI apps such as ChatGPT and Claude, and can be connected to other services through an API.
Twelve Labs CEO Lee Jae-sung said Jockey is a next-generation video-understanding AI agent that organically connects the company's core AI models.
Quick answers
What is Twelve Labs' Jockey?
Jockey is a video-understanding AI agent that lets users find and edit scenes in videos through spoken requests, including making rough cuts of highlights and isolating clips that received strong reactions.
Which apps can Jockey be used in?
Jockey can be used in AI apps such as ChatGPT and Claude, and can be integrated with services via API.
Can Jockey analyze YouTube or TikTok videos?
Yes. When a social video link from YouTube or TikTok is entered, Jockey analyzes trend and engagement metrics and recommends reference videos for promotion.